This project assumes that the effects of a given reinforcement on responding depend upon other reinforcements received either previously or concurrently. Some of the major variables manipulated include the following: The effects of performance of percentage of reinforcement, the effects on performance of a shift from one reinforcement magnitude to another (larger or smaller) reinforcement magnitude, the effects on performance of a given reinforcement magnitude when another reinforcement magnitude (larger or smaller) is currently being experienced. These and other manipulations produce results which may be explained in several different ways. Our approach, the reinforcement level view, assumes that associative effects are involved, i.e., what is learned as well as the actual amount learned is affected by the interaction of two or more reinforcement magnitudes. Our efforts are directed to some extent at confirming our approach relative to other approaches. Another concern is simply to isolate new variables which affect behavior when several reinforcement magnitudes are given together or in sequence. For example, our laboratory has shown recently that the effects of partial reinforcement on acquisition responding are larger and easier to study the longer the delay in reinforcement.